Unlock instant, AI-driven research and patent intelligence for your innovation.

A method and device for load migration

A technology of load migration and equipment, applied in the direction of network traffic/resource management, electrical components, wireless communication, etc., can solve problems such as reduced work performance

Active Publication Date: 2018-06-22
CHINA MOBILE COMM GRP CO LTD
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] Therefore, while ensuring load balancing in the LB mechanism, it is also necessary to frequently acquire and release self-selected locks, which will greatly reduce the working performance of the system.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A method and device for load migration
  • A method and device for load migration
  • A method and device for load migration

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0066] Such as figure 1 As shown in FIG. 1 , it is a schematic flowchart of a method for load migration provided by Embodiment 1 of the present invention. The method can be described as follows.

[0067] Step 101: Receive migration request information sent by a target processor.

[0068] The migration request information is used to represent the task to be processed that needs to be migrated from the source processor.

[0069] In step 101, since the current SMP environment includes multiple CPUs (Central Processing Units, central processing units), each CPU has its own scheduling domain. The Linux system searches the scheduling domain in the busiest state from the CPU scheduling domain level by level, and finds out the busiest CPU and the Run Queue corresponding to the CPU from the scheduling domain in the busiest state. The busiest CPU is used as the source processor, and some tasks need to be migrated to ensure the balance of the system's internal load. Find the least bu...

Embodiment 2

[0097] Embodiment 2 is a detailed description of each step in Embodiment 1 of the present invention.

[0098] Suppose there are N CPUs in the system, numbered as CPU 0 ,CPU 1 ,...,CPU n-1 , abbreviated as P 0 ,P 1 ,...,P n-1 ; There are N×(N-1) lock-free circular linked lists L, and L (lock-free) adopts the data structure of Ring Buffer, where there is a lock-free circular linked list between any two CPUs, for example: P i and P j The lock-free circular linked list between

[0099] Among them, each lock-free circular linked list Consists of two parts: a head pointer and tail pointer, if It means that the linked list is empty, if It means the linked list is full.

[0100] Wherein, the descriptor address of the task to be migrated is stored in the lock-free circular linked list L.

[0101] When processor P i There are tasks that need to be migrated to processor P j on, processor P i and processor P j Perform load migration corresponding to lock-free circu...

Embodiment 3

[0115] Such as figure 2 As shown, it is a schematic flowchart of a load migration method provided by Embodiment 3 of the present invention. Embodiment 3 of the present invention is an invention under the same inventive concept as Embodiment 1 to Embodiment 2 of the present invention. The method Can be as described below.

[0116] Step 201: Send migration request information to the source processor.

[0117] The migration request information is used to represent the task to be processed that needs to be migrated from the source processor.

[0118] Step 202: Obtain the descriptor address information of the task to be migrated from the intermediate table.

[0119] Wherein, the intermediate table is used by the source processor to place The source processor is determined by the rules of the task's descriptor address information.

[0120] In step 202, the intermediate table is a circular linked list, wherein the circular linked list includes first pointer information and tail ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a load migration method and device. The method comprises the steps of receiving migration request information which is sent by a target processor and used for showing a task migrated from a source processor and to be handled; determining a medium sheet which is under the coordination of the source processor and the target processor and used for storing the description character address information of the task in the source processor according to the identification information of the target processor in the received migration request message; determining the description character address information of the task to be migrated according to the capacity information of the task to be migrated and handled in the migration request information; writing the determined description character address information into the determined medium sheet so as to acquire the description character address information from the medium sheet through the target processor; obtaining the task to be migrated according to the description character address information so as to finish the migration of the load. According to the method, the medium sheet is introduced, the current self-selecting-lock based lock type load equilibrium is optimized into lock-free load equilibrium, and therefore, the performances of the whole system can be effectively improved.

Description

technical field [0001] The invention relates to the technical field of computer operation, in particular to a method and device for load migration based on a symmetrical processor SMP. Background technique [0002] In the current SMP (Symmetrical Multi-Processing, symmetric multiprocessor) environment, each CPU corresponds to a Run Queue run queue. If there is currently a task in the TASK Running (executable) state, select a Run Queue (it should be noted that only one RunQueue can be selected for the task at the same time), and add the task to the Run Queue In the run queue, so that the scheduler runs the task in the CPU corresponding to the Run Queue run queue. [0003] Since each CPU currently corresponds to a Run Queue, this allows a task that is continuously in the TASK Running (i.e. executable) state to run on the same CPU, which is conducive to data caching and improves operating efficiency. At the same time, scheduling on each CPU The tasks are all in the correspond...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): H04W28/08
CPCH04W28/08
Inventor 刘军卫钱岭
Owner CHINA MOBILE COMM GRP CO LTD